The Targetprocess MCP Server enables AI assistants to interact with your Targetprocess data through intelligent semantic operations. Beyond basic data access, it provides workflow-aware tools that understand context, suggest next steps, and adapt to your Targetprocess configuration automatically.
- Intelligent Workflows: Semantic operations that understand your work context and suggest logical next steps
- Dynamic Discovery: Automatically adapts to your Targetprocess configuration without hard-coded assumptions
- Role-Based Tools: Operations filtered by your role (developer, project manager, tester, etc.)
- Smart Error Handling: Transforms API failures into actionable guidance and learning opportunities
- Stay in Flow: Complete full workflows without switching to the Targetprocess UI
- Enterprise Ready: Handles complex schemas and millions of records with robust authentication and error handling
# Basic usage
docker run -i --rm \
-e TP_DOMAIN=your-domain.tpondemand.com \
-e TP_USERNAME=your-username \
-e TP_PASSWORD=your-password \
ghcr.io/aaronsb/apptio-target-process-mcp
# With semantic operations and strict mode (recommended for MCP clients)
docker run -i --rm \
-e TP_DOMAIN=your-domain.tpondemand.com \
-e TP_USERNAME=your-username \
-e TP_PASSWORD=your-password \
-e TP_USER_ROLE=developer \
-e TP_USER_ID=your-user-id \
-e TP_USER_EMAIL=your-email \
-e MCP_STRICT_MODE=true \
ghcr.io/aaronsb/apptio-target-process-mcp# Basic usage
TP_DOMAIN=your-domain.tpondemand.com TP_USERNAME=your-username TP_PASSWORD=your-password \
npx -y https://github.com/aaronsb/apptio-target-process-mcp.git
# With semantic operations and strict mode (recommended for MCP clients)
TP_DOMAIN=your-domain.tpondemand.com TP_USERNAME=your-username TP_PASSWORD=your-password \
TP_USER_ROLE=developer TP_USER_ID=your-user-id TP_USER_EMAIL=your-email \
MCP_STRICT_MODE=true \
npx -y https://github.com/aaronsb/apptio-target-process-mcp.gitFull installation guide → CLI usage guide →
# Quick setup for development
./scripts/dev-setup.sh
# Basic manual setup
npm install && npm run build
claude mcp add targetprocess node ./build/index.js \
-e TP_DOMAIN=your-domain.tpondemand.com \
-e TP_USERNAME=your-username \
-e TP_PASSWORD=your-password
# With semantic operations (recommended)
claude mcp add targetprocess node ./build/index.js \
-e TP_DOMAIN=your-domain.tpondemand.com \
-e TP_USERNAME=your-username \
-e TP_PASSWORD=your-password \
-e TP_USER_ROLE=developer \
-e TP_USER_ID=your-user-id \
-e TP_USER_EMAIL=your-email \
-e MCP_STRICT_MODE=trueClaude Code integration guide →
Strict Mode: For MCP clients that require clean JSON-RPC on stdio (like Claude Desktop), enable strict mode to redirect all logging to stderr:
# Environment variable
MCP_STRICT_MODE=true
# Auto-detection also works for:
# - Claude Desktop (stdio transport)
# - Non-TTY environments
# - When --stdio flag is presentSemantic Operations: Enable intelligent workflow tools with role-based filtering:
TP_USER_ROLE=developer # Options: developer, project-manager, tester
TP_USER_ID=your-user-id # For task assignments and time tracking
TP_USER_EMAIL=your-email # Identity for semantic operationsWhy Enable Semantic Operations?
- Context-Aware Tools: Get
show_my_tasks,start_working_on,complete_taskinstead of just raw API calls - Intelligent Discovery: Operations adapt to your TargetProcess configuration without hard-coded assumptions
- Workflow Guidance: Smart error handling transforms failures into actionable next steps
- Role-Based Filtering: Only see tools relevant to your role (developer, PM, tester)
# Import as a toolkit in watsonx Orchestrate
orchestrate toolkits import \
--kind mcp \
--name targetprocess \
--package-root /path/to/apptio-target-process-mcp \
--command '["node", "build/index.js"]' \
--tools "*"# Examples of what you can ask your AI assistant:
"Show me all open user stories in the mobile app project"
"Create a bug for the authentication failure on the login page"
"What's the status of our Q2 release?"
"Update the priority of story #12345 to high"
"Show me all tasks assigned to Sarah"
"Which team has the most open bugs right now?"
- Getting Started - First steps and basic usage
- Core Concepts - Understanding the key components
- Tools Reference - Detailed API documentation
- Use Cases - Common workflows and examples
- AI Integration - Setting up with Claude, ChatGPT, etc.
- Architecture - System design and implementation
- Development - Contributing and extending
- show_my_tasks: View assigned tasks with smart filtering and priority analysis
- start_working_on: Begin work on tasks with automatic state transitions
- complete_task: Mark tasks complete with integrated time logging and comments
- show_my_bugs: Analyze assigned bugs with dynamic severity categorization
- log_time: Record time with intelligent entity type discovery and validation
- add_comment: Add contextual comments with workflow-aware follow-up suggestions
- Entity Management: Create, read, update, and search Targetprocess entities
- Complex Queries: Filter items by custom fields, status, relationships, and more
- Data Discovery: Explore entity types, properties, and relationships
- Rich Includes: Retrieve related data in a single request
- Role-Based Access: Tools filtered by personality configuration (developer, PM, tester)
- Dynamic Discovery: Adapts to custom Targetprocess configurations automatically
- Error Resilience: Transforms API failures into actionable guidance
- Documentation Access: Built-in access to Targetprocess documentation
- LLM Integration: Works with Claude, ChatGPT, and other AI assistants
MIT